Hemodynamic Consequences of PEEP in Seated Neurological Patients-Implications for Paradoxical Air Embolism

ANESTH ANALG 429 1984;63:429-32

Nancy A. K. Perkins, MD, and Robert F. Bedford, MD

PERKINS NAK, BEDFORD RF. Hemodynamic 47% and right atrial pressure (RAP) increased from 3.6 +- consequences of PEEP in seated neurological patients- 0.7 SEM mm fig to 8.9 ? 0.9 SEM mm Hg (P < 0.05). implications for paradoxical air embolism. Anesth Analg Pulmona ry capillary wedge pressure (PCWP) did not in- 1984;63:429-32. crease significantly during PEEP. RAP exceeded PCWP in only two patients before PEEP, but RAP exceeded PCWP In order to better understand the hemodynamic conse- in seven patients during PEEP. Weconclude that PEEP is quences of the use of positive end-expiratory pressure (PEEP) potentially detrimental during operations in the seated po- in patients in the seated position, 22 patients undergoing sition because it not only impairs hemodynamic perform- neurosurgical operations were monitored with radial arterial ance, but might predispose patients with a probe-patent and thermistor-tipped Swan-Ganz catheters both before and foramen ovule to the risk of paradoxical air embolism. during 10-cm H20 PEEP. Significant (P < 0.05) reductions in cardiac output (15%), stroke volume (15%), and Key Words: ANESTHESIA-neurosurgical. VENTI- mean arterial pressure (24%)occurred with the introduction LATION-positive end-expiratory pressure. HEART- of PEEP, while pulmonary vascular resistance increased myocardial function. EMBOLISM, AIR-paradoxical.

Positive end-expiratory pressure (PEEP) has been ad- (9-ll), human volunteers (12), and patients with car- vocated for neurosurgical patients undergoing oper- diorespiratory disease (13-15), these studies have fo- ations in the seated position, both as a preventive cused primarily on changes in cardiac output and ven- measure to avoid air embolism (1,2) and as acute treat- tricular performance without regard to changes in ment when air embolism is suspected (3,4). In these interatrial pressure gradient. situations, it has been proposed that an increase in This study was undertaken to examine the hemo- venous pressure induced by PEEP might help to pre- dynamic impact of PEEP in anesthetized patients in vent entrainment of air into the circulation or to stop the seated position: a situation in which decreased air entrainment after air embolism has been detected. lung blood volume due to the effect of gravity (16) The use of PEEP in this setting, however, is contro- might have unanticipated effects on the cardiovas- versial, because it has been suggested that such an cular response to PEEP and in which changes in in- increase in venous pressure might promote paradox- teratrial pressure gradient due to posture may pre- ical (systemic) air embolism if the patient were one of dispose patients to paradoxical air embolism (17). the 25-35% of the population with an asymptomatic foramen ovale (5,6), thus tending to drive air bubbles from the right atrium into the left atrium and by- Methods passing the filtering function of the lungs (7,8). Al- The subjects of the study (N = 11) were informed, though the hemodynamic effects of PEEP recently consenting adult patients (ASA class I or 11) scheduled have been examined extensively in supine animals for elective cervical laminectomy at the C5-6 level. The protocol was reviewed and approved by the Hu- This study was supported in part by United States Public Health man Investigation Committee Of the University Of Vir- Services Grant T32-GM-075900-04. ginia. Anesthesia was induced with thiopental (4 Received from the Department of Anesthesiology, University mg/kg), and maintained with 70% N@ in o2and of Virginia Medical Center, Charlottesville, Virginia. Presented in part at the American Society of Anesthesiologists Annual Meeting, morphine (0.3-0.5 mg/kg Pan- Las Vegas, Nevada, October 1982. Accepted for publication De- curonium (0.1 mg/kg-- intravenously) was given to fa- cember-30, 1983. cilitate endotracheal intubation, and ventilation was Address correspondence to Dr. Bedford, Department of Anes- thesiology, University of Virginia Medical Center, Charlottesville, to maintain PaC02 between 30 and 35 mm VA 22908. Hg. Using pressure waveform control, radial and

0 1984 by the International Anesthesia Research Society ANESTH ANALG PERKINS AND BEDFORD 430 1984;63:429-32

thermistor-tipped quadruple-lumen pulmonary ar- 4PCWPl- 30 - tery catheters were placed percutaneously, and car- I diovascular pressures were transduced with Bentley PAP (mmHg) Model 800 transducers and were referenced to the I level of the 6th intercostal space anteriorly. End-tidal 0- CO, fraction (Beckman LB-2) and all cardiovascular 5- pressures were continuously recorded using a Brush I Model 440 strip-chart recorder. Placement of the pulmonary artery catheter tip in a perfused (zone 111) lung segment (16) was documented by observing a decrease in end-tidal COz fraction during each pulmonary capillary wedge pressure (PCWP) determination " (Fig. 1) (18). If a measurable decrease in end-tidal CO, I+ 1- MIN~ did not occur during balloon inflation, the position of the pulmonary artery catheter was manipulated using Figure 1. Representative strip-chart record demonstrating a de- a sterile introducer sleeve until proper placement of crease in end-tidal CO, fraction (FEKO~) during PCWP determination. Because the decrease in FEKo~ indicates a decrease in re- the catheter tip was verified (mean change in F,Heart rate (HR), mean arterial pressure (MAP), right atrial pressure Discussion (RAP), pulmonary arterial pressure (PAP), pulmo- Patients undergoing operations in the seated position nary capillary wedge pressure (PCWP), and cardiac are at a considerable hernodynamic disadvantage with output (thermodilution technique; Edwards model 9528 regard to venous air embolism. Cardiac output and cardiac output computer) were recorded at the fol- stroke volume are decreased from values in the supine lowing times: during steady-state general anesthesia state (19) and many of these patients have a right atrial with constant surgical stimulation immediately before pressure that is higher than left atrial pressure (LAP), applying 10-cm H20 PEEP to the anesthesia circuit; thus possibly placing them at risk for paradoxical air and 5 min after 10-cm H20 PEEP (Boehringer valve) embolism (17). Our results indicate that application was added to the expiratory limb of the anesthesia of 10-cm H20 PEEP in this situation causes a further circuit. reduction in cardiac output and mean arterial pres- Derived hemodynamic variables were calculated sure while simultaneously resulting in a higher right- from appropriate formulae. Values measured during to-left interatrial pressure gradient. If PEEP is used PEEP were compared to those obtained without PEEP prophylactically to prevent air embolism, it appears using Student's t-test for paired data. The incidence that small, otherwise asymptomatic, air bubbles in the of RAP > PCWP was compared before and during venous circulation might be forced through a probe- PEEP using Fisher's exact test. P < 0.05 was regarded patent foramen ovale and into the arterial circulation as significant. where coronary or cerebral damage might occur. If used as treatment of suspected air embolism, PEEP tends to mimic the hemodynamic effects of air em- Results bolism (increased pulmonary vascular resistance, right During surgery in the seated position, institution of ventricular and right atrial pressures, and decreased 10-cm PEEP caused significant reductions in MAP cardiac output) (20), and therefore, might further im- (14%), stroke volume (15%), and CO (15%)without pair cardiovascular performance at a critical juncture. a change in HR or systemic vascular resistance (Table Furthermore, because recent studies indicate delayed 1). There was a 28% increase in pulmonary artery arrival of embolized air bubbles from above the su- pressure (PAP) and a 47% increase in pulmonary vas- perior vena cava (21), rapid institution of PEEP con- cular resistance (PVR). RAP increased significantly ceivably could deliver the critical increase in interatrial with PEEP, whereas PCWP did not increase signifi- pressure gradient required to produce paradoxical air cantly, and the mean interatrial pressure gradient embolus. Although the minimal pressure gradient re- (RAP-PCWP) changed from negative to positive (Fig. quired to cause right-to-left shunting across a probe- 2). Only two patients had RAP > PCWP before PEEP, patent foramen ovale is not known, we have seen whereas seven patients had RAP > PCWP during paradoxical air embolism in a patient with only a + 5 PEEP (P < 0.05). mm Hg RAP-PCWP gradient (17), and it has been PEEP IN THE SEATED POSITION ANESTH ANALG 431 1984;63:429-32

Table 1. PEEP-induced Hemodynamic Changes in Patients in the Seated Position

0-cm HzO PEEP 10-crn H,O PEEP

Heart rate (beatsomin I) 78 2 5 78 2 5 Mean arterial pressure (mm Hg) 99.4 ? 9.4 85.6 2 3.P Mean pulmonary artery pressure (mm Hg) 14.8 2 0.7 19.0 2 1.1” Right atrial pressure (mm Hg) 3.6 2 0.9 8.9 2 O.Y Pulmonary capillary wedge pressure (mm Hg) 6.4 2 0.9 7.9 2 0.8 Cardiac output (Lamin I) 5.4 2 0.8 4.6 2 0.5” Stroke volume (ml) 69.2 8.3 59.0 2 11.2’

Systemic vascular resistance (dyne*sec*cm~ 5, 1418 ? 232 1333 2 109

Pulmonary vascular resistance (dyne*sec*cm 5, 124 2 27 192 26“

All values are mean -t scv “D < 0.05 vs 0-cm H20 PEEP

shown that a + 4 mm Hg interatrial pressure gradient pressure (3). A recent report using prophylactic PEEP can produce a 50% right-to-left shunt in patients with for neurosurgical procedures in the seated position an asymptomatic and previously unsuspected fora- found a 51% incidence of Doppler-detected air em- men ovale (22). bolism (l),considerably higher than the 36% inci- Because interpretation of the results of this study dence for similar operations at our institution without depends upon PCWP accurately reflecting LAP dur- PEEP (25). ing PEEP, we felt it was imperative to ensure that the The hemodynamic effects of 10-cm HzO PEEP in pulmonary artery catheter was always lodged in a anesthetized seated patients undergoing surgery ap- perfused (zone 111) (16) lung segment at the time PCWP pear to be similar to those previously measured in was measured. Several studies have warned against supine volunteers (12) and patients with cardiores- misinterpretation of PCWP during PEEP when the piratory disease (13-15): small reductions in cardiac catheter tip is located in a superior, nonperfused branch output, stroke volume, and mean arterial pressure of the pulmonary artery (23,24) where raised airway and a marked increase in pulmonary vascular resis- pressure is transmitted to the catheter tip. These same tance. At least three factors are thought to contribute studies, however, found that PCWP does reflect LAP to decreased cardiac output during PEEP: decreased accurately when the catheter tip is located in a per- venous return and impaired right ventricular filling fused, dependent lung segment at levels of PEEP as as a result of elevated intrathoracic pressure (26); in- high as 10-cm H20 pressure. Because we always ob- creased right ventricular afterload due to raised al- tained a measurable decrease in end-tidal COzfraction veolar pressure and compression of pulmonary cap- at the time of balloon inflation, thus indicating inter- illaries and precapillary arterioles (16); and decreased ruption of pulmonary blood flow (18), we believe that left ventricular end-diastolic volume resulting partly our measurements of PCWP accurately reflect the pa- from decreased left ventricular end-diastolic pressure tients’ LAP. Furthermore, we found that PCWP did (13,15) and partly from decreased left ventricular dis- not increase as much as RAP during PEEP, therefore tensibility (10-12). mitigating against PCWP values being falsely elevated. The impact of these hemodynamic changes is re- We chose 10-cm H20as a reasonable level of PEEP flected in our measurements of interatrial pressure to study because lower levels obviously were not ef- gradient. Increased right heart afterload would tend fective in increasing venous pressure sufficiently to to cause a relatively greater increase in RAP than LAP; prevent air entrainment in the seated position, and whereas impaired left ventricular filling and de- our pilot studies showed that higher levels (i.e., 15- creased left ventricular end-diastolic volume would cm PEEP) caused profound cardiovascular depression tend to minimize increases in LAP. The differences in seated patients with the resultant risk of cerebro- between right and left atrial pressures during PEEP vascular insufficiency due to arterial hypotension. The are more marked in the present study than those pre- fact that we could only increase mean venous pres- viously measured in supine dogs (9,lO);although spe- sure by 5 mm Hg makes us doubt that 10-cm H20 cies difference is one possible explanation, we also PEEP can reliably prevent air entrainment from open believe this discrepancy is due to the additional effect veins in an incision that is 15-20 cm above heart level. of gravity on both pulmonary blood flow and blood To our knowledge, the threshold for entrainment of volume in patients in the seated position. air into the circulation is not known, but clinical stud- We conclude that the use of PEEP therapy is not ies indicate that it might be as high as -5 cm HzO advisable during operations performed in the seated 432 ANESTH ANALG PERIUNS AND BEDFORD 1984;63:429-32

rosurgery. A comparison of various methods of detection in END EXPIRATORY PRESSURE man. Anaesthesia 1976;31:633-43. \ 9 I OcmH20 I 10 cm H20 8. Gronert GA, Messick JM JR, Cucchiara RF, Michenfelder JD. Paradoxical air embolism from a patent foramen ovale. Anes- thesiology 1979;50:548-9. 9. Qvist 1, Pontoppidan H, Wilson RS, Lowenstein E, Laver MB. Hemodynamic responses to mechanical ventilation with PEEP: the effect of hypervolemia. Anesthesiology 1975;4245-55. 10. Scharf SM, Caldini P. Ingram RH JR. Cardiovascular effects of increasing airway pressure in the dog. Am J Physiol 1977;232:H35-H43. 11. Fewell JE, Abendschein DR, Carlson CJ, Murray JF, Rapoport E. Continuous positive-pressure ventilation decreases right and left ventricular end-diastolic volumes in the dog. Circ Res Figure 2. Changes in interatrial pressure gradient (right atrial pres- 1980;46:125-32. sure [RAP]-pulmonary capillary wedge pressure (PCWP]) associ- 12. Cassidy SS, Eschenbacher WL, Robertson CH IR, Nixon JV, ated with institution of 10-cm PEEP. All values are mean ? SEM. Blomqvist G, Johnson RL JR. Cardiovascular effects of positive- *, P < 0.05 vs 0-cm HzO. pressure ventilation in normal subjects. J Appl Physiol 1979;47:453-61. 13. Rankin JS, Olsen CO, Arentzen CE, et al. The effects of airway position. It impairs cardiovascular performance when pressure on cardiac function in intact dogs and man. Circu- it is already limited due to postural changes and when lation 1982;66:108-20. 14. Prewitt RM, Oppenheimer L, Sutherland JB, Wood LDH. Ef- it might deteriorate further due to venous air embo- fects of positive end-expiratory pressure on left ventricular me- lism. It causes an increase in venous pressure that chanics in patients with hypoxemic respiratory failure. Anes- may not be sufficient to stop air entrainment from an thesiology 1981;55:409-15. incision 15-20 cm above heart level. Finally, because 16. West JB. Distribution of blood flow, In: Respiratory physiology, it increases right atrial pressure more than left atrial the essentials. Baltimore: Williams and Wilkins, 1974:42-5. pressure, PEEP would tend to promote passage of air 17. Perkins-Pearson NAK, Marshall WK, Bedford RF. Atrial pressures in the seated position: implications for paradoxical air bubbles into the systemic circulation if a patient with embolism. Anesthesiology 1982;57493-7. a probe-patent foramen ovale were to develop venous 18. Severinghaus JW, Stupfel M. Alveolar dead space as an index air embolism. of distribution of blood flow in pulmonary capillaries. J Appl Physiol 1957;10:335-48. 19. Marshall WK, Bedford RF, Miller ED. Cardiovascular responses in the seated position-impact of four anesthetic techniques. References Anesth Analg 1983;62:648-53. 1. Voorhies RM, Fraser RAR, Van Poznak A. Prevention of air 20. English JB, Westenskow D, Hodges MR, Stanley TH. Com- embolism with positive end-expiratory pressure. Neurosurgery parison of venous air embolism monitoring methods in supine 1983;12503-6. dogs. Anesthesiology 1978;48425-9. 2. Lee DS, Lichtmann MW, Weintraub HD. Effect of PEEP on air 21. Martin RW, Colley PS. Evaluation of transesophageal doppler embolism during sitting neurosurgical procedures. Anesth detection of air embolism in dogs. Anesthesiology 1983;58:117-23. Analg 1981;60:262. 22. Byrick RJ, Kolton M, Hart JT, Forbath PG. Hypoxemia follow- 3. Albin MS, Carroll RG, Maroon JC. Clinical considerations con- ing cardiopulmonary bypass. Anesthesiology 1980;53:172-4. cerning detection of venous air embolism. Neurosurgery 23. Roy R, Powers SR, Feustel PJ, Dutton RE. Pulmonary wedge 1978;3:380-4. catheterization during positive end-expiratory pressure venti- 4. Jackson PL. lntracardiac catheters unnecessary in neurosurgical lation in the dog. Anesthesiology 1977;46:385-90. anesthesia (Letter). Anesthesiology 1978;48:154. 24. Berryhill RE, Benumof JL. PEEP-induced discrepancy between 5. Edwards JE. Congenital malformations of the heart and great pulmonary artery wedge pressure and left atrial pressure. vessels: malformations of the atrial septa1 complex, In: Gould Anesthesiology 1979;51:303-8. SE, ed. Pathology of the heart. Springfield: Charles C. Thomas, 25. Bedford RF. Venous air embolism: a historical perspective. Semin 1968:262. Anesth 1983;2:169-76. 6. Thompson T, Evans W. Paradoxical embolism. Q J Med 26. Cournand A, Motley HL, Werko L, Richards DW. Physiologic 1930;23:135-50. studies of the effect of intermittent positive pressure breathing 7. Buckland RW, Manners JM. Venous air embolism during neu- on cardiac output in man. J Appl Physiol 1948;125:162-74.